How we made a text-to-speech model respond in sub-50 ms

Nari labs，Low-latency, realtime multimodal model serving, starting with speech at 50 ms time to first audio.关注
How we made a text-to-speech model respond in sub-50 ms 图片 1
添加评论
点赞收藏
点踩分享查看原文
评论
?
参与讨论